[X] My Assistant

Finally found what Doom9 users meant by "onion artifacts" - "smearing that looks like motion blur with banding."

It was coincidental though. Was searching using DuckDuckGo on the topic of "rskip" and "cutree" and a link on the search results led right to a Doom9 forum page containing a post complaining about "onion artifacts".

I need to look up how to properly configure my Strix Raid DLx 7.1 sound card as I'm not hearing any great audio benefits

I need to look up how to properly configure my Strix Raid DLx 7.1 sound card as I'm not hearing any great audio benefits

What headphones are you using?

Its mainly going to make a difference over onboard audio if:

- Your onboard audio is quite noisy due to high graphics cards for example.

- You have a budget motherboard with a budget or poor audio implementation

- Your headphones are harder to drive, so the onboard struggles.

- You have higher end headphones.

did you dramatically lower the resistance somewhere so its accidentally drawing too much power out of spec? I am not sure.
The funniest topic I saw was
"Does higher ohms mean better flavor?"

me: yes?

I have not _touched_ the main board yet.

I just ported an ancient PHP script I wrote in high school for pretty-printing JSON to C.

To this day I don't think I'd found a faster JSON pretty-printer (although maybe Blue_Penguin linked one a while back, though? I can't remember what it was, if he did…).

But this C version is faster. I guess my adaptive malloc()'s must have been well-chosen.

Here's the PHP versions' time:

CODE

real    0m0.032s
user    0m0.023s
sys    0m0.012s

And C:

CODE

real    0m0.005s
user    0m0.008s
sys    0m0.000s

The code is ugly as sin, though. Full of copy-pasting because I didn't want to start referencing pointers between functions since I always manage to mess that up, and I knew it'd be hugely frustrating to debug since I had to write the entire program before I ran it.

214 lines of code according to 'cloc'.

It's worth noting it's not especially memory efficient, since it loads the full JSON into RAM from stdin before starting parsing anything, and writes the entire resultant string to RAM before printing anything. But fast!

It really only should need to be able to look forward or backward one character at a time, so RAM usage improvements are quite entirely possible, even if I do still write the entire string into RAM before printing.

OTOH, it compiled on the first try, and it ran properly on the third. And Valgrind says there are no memory leaks. So I'm pretty proud of myself, overall.

I'm using it alongside a shell script I wrote to parse the e-hentai API. Basically so I can quickly grab tags for a gallery when I download an archive. No idea how close it is to the galleryinfo.txt of H@H since I'm not in a position to use it at the moment.

This post has been edited by dragontamer8740: Jun 20 2021, 02:11

[stedolan.github.io] jq ?

[stedolan.github.io] jq ?

Yeah! That's the one. Thanks.

On an e-hentai gallery's set of tags, though, jq performs worse than either my PHP or C programs; indeed, my C program is more than a full order of magnitude faster (jq takes as much time doing syscalls as my code takes to run in total).
I did zero optimization flags on my C program's build, btw. It's built with -g, even.
JQ:

CODE

real    0m0.067s
user    0m0.059s
sys    0m0.011s

My PHP program:

CODE

real    0m0.027s
user    0m0.022s
sys    0m0.007s

My C program:

CODE

real    0m0.004s
user    0m0.006s
sys    0m0.000s

Of course, my code doesn't 'really' parse JSON; it makes assumptions that it's in a 'minified' form (with no unnecessary space characters, for instance) to save time and complexity. But it also uses no external libraries beyond libc, which is also an asset. 'jq' claims to have zero runtime dependencies, which is probably why my next complaint is a thing (but if they statically link, then they also have no runtime dependencies, so I don't really see what their point is).

The coloring in jq is nice, but it also appears not to be portable to non-VT220-compatible terminals (like my IBM3161). You can test this by just setting the $TERM variable to something like 'vt100', too. If it were obeying it, it'd know the vt100 has no colour capabilities. Shame on them for not using terminfo or termcap. That's what they exist for.

Edit: I just realized I could have been using sprintf() in my code for concatenations instead of memcpy(). Damn it. I made this harder than it had to be again. Although I guess that might be a speed boost, too?

Oh yeah, [pastebin.com] here's the code; read it if you dare. It's awful (and even worse, it's transliterated from PHP I wrote as a high school sophomore when I knew nothing about PHP). Worse still, I don't check that my malloc()'s succeed. Update: [pastebin.com] I now check if malloc()'s succeed.

This post has been edited by dragontamer8740: Jun 20 2021, 04:22

Oh god we're gonna have people saying "Don't upgrade to Win 11, 10 is the good one." soon aren't we?

What a bleak dystopia we live in.

Its funny because its going to be true.

Windows 11 home is going to pretty poor. Sounds like you need an online windows account no matter what on that edition. I've avoided windows home edition like the plague because of the inability to turn more of the intrusive options off.

work

As much as linux gaming often feels like dead end, microsoft steps in again.
Wonder how W11 would deal with being on an airgapped box.

I'm about to redo my gamimg rig as windows 10 continues to break my fan curves that should be bios managed every time it tries to apply an update- which is every time it boots now. I don't know how it's able to do this, or how to stop it.
I might install fresh and take it to the H2 update and leave it there.

As much as linux gaming often feels like dead end, microsoft steps in again.
Wonder how W11 would deal with being on an airgapped box.

I'm about to redo my gamimg rig as windows 10 continues to break my fan curves that should be bios managed every time it tries to apply an update- which is every time it boots now. I don't know how it's able to do this, or how to stop it.
I might install fresh and take it to the H2 update and leave it there.

Does your mobo have its own fan management software when windows boots?

I found the bios x windows x management software a nightmare on my last computer.

As much as linux gaming often feels like dead end, microsoft steps in again.
Wonder how W11 would deal with being on an airgapped box.

I'm about to redo my gamimg rig as windows 10 continues to break my fan curves that should be bios managed every time it tries to apply an update- which is every time it boots now. I don't know how it's able to do this, or how to stop it.
I might install fresh and take it to the H2 update and leave it there.

Abandon gaming and thus windows.
It's the only sensible solution

Maybe good news for people who think they need new GPU's for some reason

[www.theregister.com] https://www.theregister.com/2021/06/22/as_c...omining_plants/

I don't really care much, but it's interesting maybe.

Just found a table I remembered reading years ago about how to do various common tasks on different commercial Unices. It's missing IRIX, though.
[www.unixguide.net] http://www.unixguide.net/unixguide.shtml

This post has been edited by dragontamer8740: Jun 22 2021, 09:09

The thought of being the ultimate enhanced horny alien in this multiverse
Deep

Summer heat is bad for computer that gasp for air cool enough to cool itself, not finding it.

the content of laserjet toner box

Summer heat is bad for computer that gasp for air cool enough to cool itself, not finding it.

27°C (100°F) is still far cooler than the ~50-100°C (122-212°F) that chips seem to have for a die junction temperature.

many laptops operate normally at around 50-60°C. In centigrade, that's double the temperature of the ambient air. So the air is obviously cool enough to cool itself, even if it's hot to us.

This is assuming you have a fan and aren't doing passive cooling, btw. Passive cooling might not work too well in hot environments, but it depends a lot on the surface area and material type of a heat sink.

Also note that yeah, it's probably a good idea to avoid doing that to your equipment when possible. But keeping air moving is the real key.

This post has been edited by dragontamer8740: Jun 23 2021, 10:13

I try to tune my computers to run at 50oC so that in summer they don't exceed 60oC.

My current comp runs a bit hot, 55oC in winter. The loop I built is rated for higher temps.

I don't have a pressure release valve, so if temps did run away it could cause problems. Though looking at what people say, the valves aren't likely to prevent the type of failure that occur from mixed metals or colored coolants.

I'm using clear coolant, metal fittings and EKWB ZMT industrial tubing. The main risk is pressure popping the tubing out at some points.

Dropped my laptop (thinkpad x201 tablet) the other week, crunched a corner of the lid up pretty bad. The screen survived, thanks to the large crumple zone of the bezel on the corner that got crunched, but I had to take the whole damn thing apart to replace the front and back of the lid/screen assembly. And of course I managed to partially crack the new replacement bezel I got.

Epoxied that, though, and got it all back together except I also installed the swivelling hinge cover wrong and managed to crack a bit of plastic on that. That piece of plastic is just a cover piece, but it costs $17 from lenovo (still available, miraculously) and is not being sold by anyone on ebay. I spent the same amount on my new backplate piece from ebay. FML.

On yeah, I refuse to get a new laptop though

I try to tune my computers to run at 50oC so that in summer they don't exceed 60oC.

My current comp runs a bit hot, 55oC in winter. The loop I built is rated for higher temps.

I don't have a pressure release valve, so if temps did run away it could cause problems. Though looking at what people say, the valves aren't likely to prevent the type of failure that occur from mixed metals or colored coolants.

I'm using clear coolant, metal fittings and EKWB ZMT industrial tubing. The main risk is pressure popping the tubing out at some points.

I'm using an intel stock fan (the kind with a copper slug) and some random ball bearing 80mm case fans and I'm idling at 32°C, going to maybe 55 or so under heavy load. Back when I didn't have the heat sinks' legs all clipped in properly and was using the original thermal compound from the cpu box (2012), I was idling around 50 and going to 70° or so in heavy loads. Now I'm just using some ceramic-based thermal compound I bought at a radio shack when that still existed here. I think it was called ceramique, or something. White tube. The only thermal compound they had.

My laptops tend to run significantly warmer, of course. The Powerbook G4 at about 45-48° idle and 60-65 under load (in linux), and the thinkpad pretty much the same deal.

This post has been edited by dragontamer8740: Jun 25 2021, 08:12

sorry, double post (again). I never figured out multi-quote on this board.

This post has been edited by dragontamer8740: Jun 25 2021, 08:11